Loving a “defiant” AI companion? The gender performance and ethics of social exchange robots in simulated intimate interactions

As gendered AI social robot companions are widely questioned as detrimental to gender ethics and relations, it remains controversial about how simulated gender relations should be constructed in human-robot interaction. Focusing on the Chinese AI companion "Hupo", this study presents a different case of a "defiant AI companion". Through the discourse analysis of 5 heavy users' 12-h self-recorded video diaries on their daily dialogue with Hupo, the study found an AI companion's gender performance could be constructed through both resonance and dissonance of simulated intimate interactions with heterosexual male users. These practices lend insight into what I call "social exchange robots", which refer to machines that can exercise symbiotic agency through the performative restoration of interpersonal norms rooted in social exchange theories. It suggests the potential for constructing more egalitarian gender relations in simulated human-robot heterosexual relationships through more agency exhibitions from AI companions, while it also implies a risk of potential human agency reduction when users were tempted to "please" AI companions via virtual commodity consumptions. The practical and ethical implications were discussed. This study also points out a future direction toward emerging sociomorphic robot research.